Enhancing Histopathology Image Generation With Diffusion Generative Models: A Comprehensive Study

Abstract

  The field of histopathology faces significant challenges due to the limited availability of data, which is often not publicly accessible due to privacy issues. The scarcity of publicly available high-quality datasets also hampers the development and training of effective deep learning models. Generative Adversarial Networks (GANs) have previously attempted to address these issues by creating synthetic data but suffer mode collapse, which reduces their effectiveness and reliability. This study explores Diffusion Generative Models (DGMs) as a unique and robust alternative for generating synthetic pathology images.

The primary objective of this study is to compare various Diffusion Generative Models (DGMs) and methods in medical imaging. Specifically, we examine the Denoising Diffusion Probabilistic Model (DDPM) and the Latent Diffusion Model (LDM), along with other generative sampling choices. Both models demonstrated the ability to generate realistic histopathological images. We also investigated DGMs from a unique perspective by generating various patch sizes, demonstrating that DGMs effectively learn patch resolution.

Additionally, we analyze the application of DGMs across different fields of view (FOV) extracted from whole slide images of the Kingston General Hospital (KGH) dataset to investigate their impact after generating synthetic data. Our comparison includes the performance of DGMs with FOVs of 224 and 336. Our experiments showed that a 336 FOV achieved better performance with an FID score of 18.45 compared to 19.08 for the 224 FOV. The 336 FOV yielded a classification accuracy of 91.15%, surpassing the real dataset’s 80.06% and the combined dataset’s 88.21%.

In computational pathology, generative models can enhance data sharing and augmentation, improving the accuracy of deep learning classifiers, and assisting in the cancer diagnosis workflow, thereby advancing digital pathology. Our research findings confirm this and set the stage for future developments in the pathology workflow.